OpenAI just made it dramatically easier to run shopping ads inside ChatGPT. Reported on June 2, 2026 (with first detailed coverage on May 12, 2026), ChatGPT Ads Manager now has a "product feed" campaign type. You upload a full catalog, the same feed you already use for Google Shopping, set eligibility filters, and the system auto-generates sponsored placements from your product titles, images, pricing, and attributes. No building a campaign per product.
For Shopify merchants already running Google Shopping, this is close to free distribution into a fast-growing discovery surface. The feed you maintain today is most of the work. Here's how to set it up, what the 100-product gate means, and the feed hygiene that decides whether you actually get surfaced.
Watch
OpenAI explains the thinking behind ads in ChatGPT:
What shipped, and why it's a big deal for paid search
The mechanics are deliberately low-friction. Retailers upload their full catalog, set eligibility filters, and ChatGPT auto-generates sponsored placements from the feed's titles, images, pricing, and attributes. It supports up to roughly 1 million SKUs per advertiser, so catalog size won't be your bottleneck. Ads appear labeled "Sponsored" below ChatGPT answers, and Criteo is named as an early ad-tech partner.
The strategic point: this isn't a separate creative production effort. It's an extension of the Google Shopping feed discipline you already (should) have. That makes it a natural line item in a paid-search budget rather than a new channel you build from zero. For the broader playbook on advertising inside ChatGPT, see our ChatGPT shopping ads 2026 e-commerce playbook.
Step 1: Reuse your existing Google Shopping feed
The single biggest benefit is that ChatGPT's product feed accepts the same catalog feed used for Google Shopping. If you run Google Shopping on Shopify, you already produce this file. Point it at ChatGPT Ads Manager and you're most of the way there.
Before you do, run a feed quality pass, because the same fixes lift both channels. Our Google Shopping feed optimization checklist for 2026 is the right starting point. Clean the feed once; benefit twice.
Step 2: Get the feed hygiene right (this decides everything)
Because ChatGPT generates placements directly from your attributes, feed quality is the whole ballgame. The fields that matter most:
- Titles. Front-load the attributes a shopper would actually say to ChatGPT (brand, product type, key attribute). Vague titles get buried.
- Images. Clean, accurate primary images. The placement is visual.
- Pricing. Accurate and current. Mismatches between feed and store erode trust and can trigger review issues.
- Attributes. Complete and specific. Color, size, material, category, and any spec a query might hinge on. ChatGPT uses these to decide relevance.
Because ChatGPT increasingly favors brands and products it can confidently verify, the same structured-data discipline that wins organic AI visibility helps here too. Our Shopify ChatGPT shopping optimization guide covers that overlap.
Step 3: Pass the 100-product gate
New advertising partners must submit a sample of roughly 100 products before OpenAI enables the full catalog. Treat this as a quality checkpoint, not a formality. Send your 100 best-structured SKUs, the ones with complete titles, clean images, accurate pricing, and full attributes, so the sample sails through and your full feed gets switched on.
A practical move: pick 100 high-margin, high-demand products and over-invest in their feed quality first. They'll both clear the gate and be your strongest performers once live.
Step 4: Set eligibility filters and fit it into the budget
Eligibility filters let you control which products are in play. Use them to exclude low-margin, out-of-stock, or poorly-attributed items so you're only spending against products that are ready to convert.
On budgeting, think of ChatGPT product feed as an extension of paid search, not a replacement for it. Start with a controlled portion of spend, measure incremental orders, and scale into what works. Because it shares feed plumbing with Google Shopping, the operational overhead is low, which is exactly why it deserves a test allocation. For how AI shopping traffic actually converts once it lands, see converting AI shopping traffic for e-commerce.
What to do this week
- Pull the catalog feed you already use for Google Shopping and run it through a quality pass on titles, images, pricing, and attributes.
- Build a clean 100-product sample from your best high-margin SKUs to clear the OpenAI gate on the first try.
- Set eligibility filters to exclude out-of-stock, low-margin, and weak-attribute products.
- Allocate a controlled test budget, measure incremental orders, and scale into winners.
If you're not on Shopify yet and want the cleanest path to a feed that serves both Google Shopping and ChatGPT, you can launch your store on Shopify and inherit that catalog plumbing from day one.
The brands that win here aren't the ones with the biggest catalogs, they're the ones whose feeds are clean enough that ChatGPT can confidently match a product to a shopper's question. Fix the feed once, clear the gate, and you've added a real paid-search surface for almost no incremental creative work.